Post-remission measurable residual disease directs treatment choice and improves outcomes for patients with intermediate-risk acute myeloid leukemia in CR1

Objectives This study retrospectively investigated in which cycle measurable residual disease (MRD) is associated with prognosis in patients in first complete remission (CR1) of intermediate-risk acute myeloid leukemia (AML). Methods The study enrolled 235 younger patients with intermediate-risk AML. MRD was evaluated by multiparameter flow cytometry after the 1st, 2nd, and 3rd chemotherapy cycles (MRD1–3, respectively). Results No significant association was detected after the 1st and 2nd cycles. However, the 5-year incidence of relapse was higher in the MRD3-positive group (n = 99) than in the negative group (n = 136) (48.7% vs. 13.7%, P = 0.005), while 5-year disease-free survival (DFS) and overall survival (OS) were lower in the MRD3-positive group than in the negative group (43.2% vs. 81.0% and 45.4% vs. 84.1%; P = 0.003 and 0.005, respectively). Allogeneic hematopoietic stem cell transplantation led to a lower 5-year relapse, and higher DFS and OS rates than chemotherapy in the MRD3-positive group (22.3% vs. 71.5%, 65.9% vs. 23.0%, and 67.1% vs. 23.9%; P < 0.001, 0.002, and 0.022, respectively), but did not affect the MRD-negative group. Conclusions MRD3 could serve as an indicator for post-remission treatment choice and help improve outcomes for intermediate-risk AML in CR1. Supplementary Information The online version contains supplementary material available at 10.1007/s12185-022-03441-6.


Introduction
For adult patients with acute myeloid leukemia (AML), complete remission (CR) rates are approximately 70-80%. However, approximately 60% of these patients experience a relapse in their first CR (CR1) despite intensive chemotherapy [1]. To improve the leukemia-free survival of patients with AML, distinct therapeutic strategies for postremission consolidation have been developed. Treatment decisions are based on cytogenetic risk profiles, and, recently, on the measurable residual disease (MRD) [2,3].
In the treatment of intermediate-risk AML (IR-AML), post-remission relapse is a major concern; but no consensus on post-remission treatment exists. Some reports suggest that high-dose combined chemotherapy is a viable option for patients with IR-AML [4,5]. Other studies have indicated that allogeneic hematopoietic stem cell transplantation (allo-HSCT) was associated with better survival compared with chemotherapy when applied as post-remission therapy (PRT) [6][7][8][9]. Interestingly, the HOVON/SAKKAML 42A study mentioned that MRD plays a crucial role in deciding which PRT options would be the most effective [10]. Another study, which investigated standard-risk AML in CR1 according to the second cycle MRD status, could not indicate whether allo-HSCT was significantly useful in MRD-positive cases (95% confidence interval (CI) 0.31-1.69) [11]; in the IR-AML subgroup following the GIMEMA AML1310 protocol [12], no significant differences in survival between the second cycle (one induction and one consolidation cycle) MRD-positive receiving allo-HSCT and MRD-negative receiving auto-HSCT subgroups were found. However, the selection of strategy (including allo-HSCT, auto-HSCT, and chemotherapy) for IR-AML, and the MRD cycle for deciding the therapeutic strategy remains controversial. Thus, which specific cycles of MRD are closely associated with prognosis should be clarified; this would be helpful in determining post-remission treatment strategies for IR-AML in CR1. We conducted a retrospective study examining younger adults with de novo IR-AML in CR1 who underwent either chemotherapy, auto-HSCT, or allo-HSCT as consolidation therapies in our centers.

Patients
Younger patients in our centers (the First Affiliated Hospital of Zhengzhou University, the First Affiliated Hospital of Xinxiang Medical University, and Huaihe Hospital of Henan University) who had a confirmed diagnosis of IR-AML between January 2015 and December 2018 were enrolled if they met the following criteria: (1) newly diagnosed with de novo IR-AML, (2) between the ages of 14 and 60 years, (3) received ≥3 cycles of chemotherapy and in CR1. The study is in accordance with the principles of the Declaration of Helsinki and was approved by the ethical committee of the First Affiliated Hospital of Zhengzhou University, the First Affiliated Hospital of Xinxiang Medical University, and Huaihe Hospital of Henan University.

Treatment response definitions
CR was defined as follows: <5% BM blasts; absence of blasts with Auer rods; absence of extramedullary disease; absolute neutrophil count >1.0 × 10 9 /L; platelet (PLT) count >100 × 10 9 /L, and independence of red cell transfusions. CRi corresponded to all CR criteria except for neutropenia (<1.0 × 10 9 /L) or thrombocytopenia (<100 × 10 9 /L). PR was defined as 5% <BM blasts <25% and a decrease in the pre-treatment bone marrow blast by at least 50%. NR was defined as BM blasts ≥25%; relapse was a recurrence of ≥5% BM blasts, the reappearance of blasts in Post-remission measurable residual disease directs treatment choice and improves outcomes… the blood, or the development of extramedullary disease, and non-CR included NR and PR as described previously [13].

Monitoring and definition of MRD
BM samples were collected to monitor MRD using 8-color multiparameter flow cytometry (MFC) after each course of chemotherapy (induction and each cycle of consolidation). MRD was detected using a panel of eight antibody combinations [14]. LAIP (leukemia-associated immunophenotype) defined at AML diagnosis was used for MFC-MRD detection. The different-from-normal immunophenotype was performed to monitor MFC-MRD when LAIP was not available at diagnosis. The sensitivity of MFC-MRD detection was 0.1%, and any level of measurable MRD ≥0.1% was considered positive, and ≤0.1% was negative. MRD1-3 refers to the MRD after the 1st, 2nd, and 3rd cycles of chemotherapy, respectively.

Endpoints and statistical methods
Cumulative incidence of relapse (CIR) was the primary endpoint of the study. Secondary endpoints included diseasefree survival (DFS) and overall survival (OS), as estimated using the Kaplan-Meier method and compared by the logrank test. Both CIR and DFS were measured from the date of achievement of the first remission, and OS was measured from the date of diagnosis. The χ2 analyzed categorical variables or Fisher's exact test, and differences between distributions of continuous variables were evaluated using the Mann-Whitney U test. The area under the receiver operating curve constructed from logistic regression analysis was used to present the associations between MRD and relapse. Based on the cut-off value for relapse, MRD was defined as negative or positive.
The Cox proportional hazards model was applied to evaluate the association between patient characteristics and outcomes in multivariate analysis. Variables associated with relapse or survival with P-values less than 0.10 in univariate analysis or variables (e.g., age, high white blood cell (WBC), cycles required to achieve CR ≥2) known to influence outcomes were included in the final models. P-values of <0.05 whereas considered statistically significant. SPSS 22.0 (Mathsoft, Seattle, WA), GraphPad Prism 6 (GraphPad Software Inc., La Jolla, CA), and R software (http:// cran.Rproje ct. org) were used for all data analysis.

Patient characteristics
From January 2015 to December 2018, 337 consecutive younger adult patients were diagnosed with IR-AML in our centers (Fig. 1). Of the 337 patients, 21 lacked complete cytogenetic data at diagnosis, and 25 were not treated; both were excluded from the study. Of the remaining 291 patients, 32 were excluded from this study due to death (n = 19) or loss to follow-up (n = 13) during the two cycles of induction chemotherapy. Of the remaining 259 patients, those with non-CR (n = 12), loss to follow-up (n = 4), missing MRD measures (n = 6), or mortality (n = 2) were excluded after another cycle of chemotherapy consolidation. The remaining 235 patients were enrolled.
After The characteristics of the patients are summarized in Table 1. The median age of the patients was 36 years, and the WBC count at diagnosis was ≥100 × 10 9 /L in 29 patients. After the first induction cycle, 181 patients achieved CR/CRi, including 51 MRD-negative cases (MRD1 <0.1%); after the second induction cycle, 215 achieved CR/CRi including 82 MRD2-negative cases; and after the third cycle, 133 (56.6%) patients were MRD3-negative. The median follow-up period for this cohort was 18 (3-72) months.

Interaction between MRD and beneficial effects of HSCT in CR1
In this study, 42 patients who were MRD3 negative underwent auto-HSCT in CR1. Among the other patients, consolidation chemotherapy and allo-HSCT were administered to 114 and 79 cases respectively. The interaction between MRD1-3 and HSCT and its effects on patient outcomes were assessed. Among chemotherapy, allo-HSCT, and auto-HSCT subgroups of MRD1-positive and -negative patients, the estimated CIR, and survival are shown in the Supplementary Material; the 5-year cumulative OS was not significantly different (P > 0.05, Figure S2). Among chemotherapy, allo-HSCT, and auto-HSCT subgroups of MRD2-positive and -negative patients, CIR and survival are shown in the Supplementary Material; the estimated 5-year OS was not significantly different (P > 0.05, Figure S2). The effects of MRD3 and HSCT on patient outcome were also evaluated (Fig. 3). The results indicated that outcomes were improved by allo-HSCT (n = 47) compared with consolidation chemotherapy (n = 55) for patients with MRD3 positivity. For MRD3-negative patients, no significant difference between different therapies was found in 5-year cumulative relapse (13.9% vs. 9.5% vs. 17.3% in chemotherapy, allo-HSCT, and auto-HSCT subgroups, respectively; P = 0.469, Fig. 3). Similarly, there was no statistically significant difference in the estimated 5-year DFS and OS rates among the three subgroups (DFS: 78.1% vs. 83.0% vs. 82.7%; OS: 79.0% vs. 87.7% vs. 85.6%; P = 0.805 and 0.940; respectively, Fig. 3). However, for MRD3-positive patients, the 5-year CIR was lower in the allo-HSCT than in the chemotherapy subgroup (22.3% vs. 71.5%, P < 0.001, Fig. 3); the 5-year cumulative DFS and OS were higher in the allo-HSCT than in the chemotherapy subgroup (65.9% vs. 23.0%, and 67.1% vs. 23.9%; P = 0.002 and 0.022, respectively, Fig. 3). Post-remission measurable residual disease directs treatment choice and improves outcomes…

Discussion
AML is a heterogeneous clonal disease with dismal survival, largely due to high relapse incidence [20]. Although more than 70% of patients will achieve CR1 after induction chemotherapy, many may relapse [21]. In previous studies, MRD status was proven to be an independent prognostic predictor for oncologic outcomes [10,22,23]. The precise timepoint at which MRD is associated with prognosis remains controversial. In this study, we demonstrated that MRD3-positive patients (i.e., after the third chemotherapy cycle, two inductions and one consolidation) were more likely to relapse compared to those with MRD3 negativity in the population of younger patients with de novo IR-AML. Additionally, MRD3-positive patients had lower 5-year DFS and OS. However, these unfavorable outcomes were ameliorated more effectively by allo-HSCT than chemotherapy. Therefore, our results indicated that MRD3 (after the third cycle of chemotherapy) was an appropriate indicator of PRT choice for IR-AML in CR1.
For patients with IR-AML who have achieved CR after induction therapy, additional PRT is an essential component for maintaining a disease-free status. Three PRT alternative strategies, including conventional chemotherapy, auto-HSCT, and allo-HSCT, could be chosen as consolidation therapy for post-induction. The centers have different choices that depend on factors such as patient characteristics, availability of a donor, and doctors' discretion [24][25][26][27]. Auto-HSCT has higher availability and a lower rate of nonrelapse mortality than allo-HSCT. However, it is associated with a higher relapse rate, due to a lack of a graft-versusleukemia effect [28]. Allo-HSCT following a myeloablative conditioning regimen has been considered an effective option for reducing the risk of relapse for patients with AML in CR1, particularly in younger patients with MRD-positive status. These patients should preferentially be offered myeloablative rather than reduced-intensity conditioning allo-HSCT [29][30][31]. Nevertheless, the timepoint at which MRD is most closely associated with prognosis remains controversial. A few studies showed that MRD after cycle 1 or 2 was associated with prognosis, while another study suggested that MRD was an independent prognostic factor after several cycles by next-generation sequencing (such as pre-transplantation) [10,11,32].
Several factors influence MRD status and the subsequent outcomes of patients with IR-AML, including the chemotherapy intensity, infection, HSCT application, and survival [33][34][35]. Recently, a study indicates that highintensity relative to low-intensity chemotherapy results in higher rates of MRD-negative cases [35]; another study indicates that induction therapy of IDAC with daunorubicin and omacetaxine mepesuccinate increases survival in patients with AML aged 15-55 years relative to conventional-dose cytarabine [36]. Luca Maurillo reported that MRD negativity is not less in the standard-dose group as compared to high-dose cytarabine after both induction and consolidation [37]. Three doses of cytarabine consolidation (1.5 g/m2, 2 g/m 2 , and high-dose cytarabine (HiDAC) 3 g/ m 2 ) were administered to intermediate-and poor-risk AML patients, and showed no significant differences in DFS and OS; however; septic shock was significantly higher after HiDAC 3 g/m 2 administration as compared to the IDAC regimen [38]. The present study indicated that MRD direction and HSCT selection in the PRT were vital for improving the survival of IR-AML patients.
In our study, we explored the role of MRD1-3 in directing the selection of PRT for patients with de novo IR-AML in CR1, although the selection for transplant was dependent on MRD status and donor availability. As demonstrated by our results for patients who were MRD3 negative, outcomes did not differ among high-dose chemotherapy, auto-HSCT, and allo-HSCT treatment. Venditti et al. [12] also reported that auto-HSCT exerts an effect similar to that of allo-HSCT on patients with IR-AML and MRD negativity. However, for our patients with MRD3 positivity, allo-HSCT yielded a better outcome than consolidation chemotherapy; it significantly decreased relapse rates and increased DFS and OS. Additionally, different outcomes were not found among the MSD, MUD, and HID transplant subgroups, although the sample sizes were limited. This agrees with results from recent studies, which have also indicated that clinical outcomes for HID are equivalent to MSD and MUD transplantation [39][40][41][42]. Therefore, consolidation chemotherapy and auto-HSCT could be considered a valuable alternative after induction therapy, which may be preferred for patients with IR-AML and MRD3 negativity in the absence of a suitably matched donor. In the presence of low resources, allo-HSCT should be reserved for patients with IR-AML and MRD3 positivity.
Our results also demonstrated that MRD3, but not MRD1 or MRD2 was closely related to relapse and survival. MRD3 could be interpreted as insufficient 'depth of remission' from the first two chemotherapy cycles, and patients should undergo or prepare for allo-HSCT to prevent relapse and protect organs at the time of the fourth chemotherapy cycle. We also found that a 0.1% cut-off value for MRD was appropriate to discriminate between patients with significantly different CIRs. In addition, a recent study by Paiva et al. [43] indicated that a cut-off value of 0.1% was suitable to discriminate patients with different relapse rates, although other reports showed different values [44,45].
This study had some limitations. First, our study was not randomized, and the number of patients was not large. Second, the possibility of treatment bias exists, as patients with MRD positivity may be less likely to receive auto-HSCT due to the high rate of relapse [46]. In addition, older patients were not included in this study because of their inability to receive HSCT. Therefore, larger, prospective randomized clinical trials are required to validate, modify, or disprove our results with more certainty.
In summary, our results indicate that MRD3, rather than MRD1 or MRD2, could be used, not only as a potent predictor of outcomes, but also as an indicator of optimal subsequent treatment strategies for patients with de novo IR-AML in CR1. The findings of this study also suggest that patients with IR-AML could be further categorized based on MRD3 status, which could improve personalized PRT.

Declarations
Conflict of interest The authors declare that they have no conflicts of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.